A64 OC Data

Status
Not open for further replies.
jinu117 said:
3500+ Newcastle @ 3.05ghz with 1.85 actual v-core
or
3500+ Newcastle @ 2.98ghz with 1.7 actual v-core
Both setting Prime 95 stable. Still haven't pushed how high bench setting I can get but preliminary around 3.15ghz... :p Gotta wait for my 2 sets of TCCD (one coming in today and one tomorrow :p) for pushing it harder.
Cooled by prometeia Mach II modified by baker 18 priming around -46c evap on 26c ambient ;P
As for mosfet, it doesn't get too hot considering the power I am pushing through mobo... around 58-60c... there are some air coming from fan situated above memory. Think neo2 just have used decent parts on it's PWM circuit at least to make it maybe more efficient... or the heat is getting absorbed through CPU :p
Click to expand...

that's cool
any pixs available?
 
Darth_Fluffy said:
So, is that why those very same guys on Xtremesystems killed a bunch of A64s, phase changers, or no phase changers?
Click to expand...

Running high Vcore with phase change for benches doesn't really matter. Especially with cascades in the -90 to -110 C range :)

Most of the A64 memory controller deaths were on the Shuttle AN50R...no such problem has occurred on a mass scale since so it makes sense that it was a problem unique to the Shuttle board.

Plus, there's a post that I read on the DFI forum stating that Vdimm does not affect the memory controller, only the Vcore since the memory speed is based off the cpu speed using a divider in K8 architecture.
 
DaveX said:
Running high Vcore with phase change for benches doesn't really matter. Especially with cascades in the -90 to -110 C range :)

Most of the A64 memory controller deaths were on the Shuttle AN50R...no such problem has occurred on a mass scale since so it makes sense that it was a problem unique to the Shuttle board.

Plus, there's a post that I read on the DFI forum stating that Vdimm does not affect the memory controller, only the Vcore since the memory speed is based off the cpu speed using a divider in K8 architecture.
Click to expand...
Oh, yes, AN50R. Who knows why they all decided to use that rogue board?
 
Darth_Fluffy said:
Oh, yes, AN50R. Who knows why they all decided to use that rogue board?
Click to expand...

Because at the time it was one of the only boards with an AGP lock.
 
I've found my best results @ 2380 (265fsb) with the mem running at 220mhz (5:6 ratio), I can run the fsb up to 280 but the hit I take on having the mem run at 187 (2:3 ratio) seems to lose more than i gain. Ive been running pcmark04 to gauge my results 49xx at 265 vs. 48xx at 280. I havent pushed the cpu over 1.475 volts yet to see if it will go higher to bring the mem closer to 200mhz at 2:3 ratio.

quick questions, what is the highest - safe voltage these processors can take?
also I had to bump up my mem voltage a hair, even at 280fsb w/2:3 ratio=187 - if its not running at 200mhz why was that necessary?
 
powersupply?? when you increase the voltage it can strain the other lines under load. Just a shot in the dark though.
 
I also noticed having to bump memory voltage to keep it stable even when it was UNDERCLOCKED from it's stock speeds....
What CAS latency are you running? I'm at 2-2-2-5-1T, perhaps the mobo's are having some problems keeping those tight timings... i'm pretty sure my memory is holding me back from 2500-2600mhz OC as CPU speeds stay steady at less than 35c... machine is stable but errors out of prime95 :(

I'm running a zippy500w w/ a NU (single molex) 6800 and 2 hd's so i don't think power is the problem :/
Anyone know which rails are taxed by memory the most? 3.3v perhaps?

Zoramac said:
I've found my best results @ 2380 (265fsb) with the mem running at 220mhz (5:6 ratio), I can run the fsb up to 280 but the hit I take on having the mem run at 187 (2:3 ratio) seems to lose more than i gain. Ive been running pcmark04 to gauge my results 49xx at 265 vs. 48xx at 280. I havent pushed the cpu over 1.475 volts yet to see if it will go higher to bring the mem closer to 200mhz at 2:3 ratio.

quick questions, what is the highest - safe voltage these processors can take?
also I had to bump up my mem voltage a hair, even at 280fsb w/2:3 ratio=187 - if its not running at 200mhz why was that necessary?
Click to expand...
 
groebuck said:
powersupply?? when you increase the voltage it can strain the other lines under load. Just a shot in the dark though.
Click to expand...

Thats possible, but I did just install a brand new PS two days ago with the exact same results. I went from a 420w raidmax to a 480w antec trueblue (although the raidmax seemed to have smaller voltage fluctuations). I set my mem timings to 2.5-3-3-10 2T (what is the difference between 1T and 2T?)
 
280 FSB with a 5:6 ratio would only mean 13mhz more for the memory over the 220 it'd run at with the FSB at 265... The mem can't handle it? Should be able to keep reasonable timings at 233 still.

Many newer sticks of performance memory (and even older ones like BH5, to a much higher degree) just like more voltage to reach optimum settings than the average generic stick does, I wouldn't sweat having to raise the V on them. They even come rated for 2.7/2.8v most of the time.

You should be able to do 1T by the way, it'll be a major boost in performance, much more than any of the other timings... No need to go 2T until very high memory clocks (or unless you're using multiple 1GB sticks and filling all slots).
 
AMD specs K8 to 1.65V I believe. Running that 24/7 on air should be fine, as long as the temps are not like 60 C ;)
 
60 idle or load? 60 load isn't high.. Mine goes into high 70's all the time. Of course, I'll never know exactly how much; my board is a lying communist.
 
Yes. Well, I'm going watercooling soon, so I can break 2.5GHz. And I'm also getting one of the high-end cards. And a new board.
 
Athlon 64 3000+ @ 2.500.02 MHz(watercooled)
Nvidia 6800GT @ 401/1.0 (watercooled)
OCZ 1GB (2x512) Pc4000
(CPU-z Version 1.23.2
timings:
Frequency - 208.3
FSB:DRAM - CPU/12
2.5,4,4,6,10,16
fsb 250x2,10X multiplier)
themaltake 480 PSU
 
Sgt.Corleone said:
Athlon 64 3000+ @ 2.500.02 MHz(watercooled)
Nvidia 6800GT @ 401/1.0 (watercooled)
OCZ 1GB (2x512) Pc4000
(CPU-z Version 1.23.2
timings:
Frequency - 208.3
FSB:DRAM - CPU/12
2.5,4,4,6,10,16
fsb 250x2,10X multiplier)
themaltake 480 PSU
Click to expand...


is this a 939 or 754? I thought all the 939s were multiplier locked above thier stock value? or is there something special you did to get the multiplier to 10x if its a 939
 
Zoramac said:
is this a 939 or 754? I thought all the 939s were multiplier locked above thier stock value? or is there something special you did to get the multiplier to 10x if its a 939
Click to expand...
look at his sig, k8n neo. it's a socket 754 board.
 
Sgt.Corleone said:
Athlon 64 3000+ @ 2.500.02 MHz(watercooled)
Nvidia 6800GT @ 401/1.0 (watercooled)
OCZ 1GB (2x512) Pc4000
(CPU-z Version 1.23.2
timings:
Frequency - 208.3
FSB:DRAM - CPU/12
2.5,4,4,6,10,16
fsb 250x2,10X multiplier)
themaltake 480 PSU
Click to expand...



that info is not possible.
if its a 3000+ then
#1 - the mulitplier is infact 9 not 10. or is it 8 ?
#2 - a FSB setting of 250mhz does NOT make 2.5ghz - my 3200+ CG runs 250mhz at 2.5ghz. there for your 3000+ is only running,, what ?? 2.3ghz ??

so recheck which chip you have cuz if its a 10x multiplier then you have a 3200+ not a 3000+
 
what makes me so mad about my overclock is that im stuck at 2.5ghz.

nothing i do allows for a higer clockrate.

i have NOT however dropped the multiplier down to 9 or 8 and check to see what i can crank the fsb up to.
simply because the mofo's at asus only allow for vdimm voltages of 2.7v MAX,, thats the most stupid thing ive seen a mobo maker do and most suprising is that asus did it.. those idiots should now better.

anyone know if the asus K8N-E is doing anybetter than what my sig says ?
i am thinking of getting one of the ocz ddr boosters but i fear i may waste my money if i cant get a better overclock.
 
v_lestat said:
that info is not possible.
if its a 3000+ then
#1 - the mulitplier is infact 9 not 10. or is it 8 ?
#2 - a FSB setting of 250mhz does NOT make 2.5ghz - my 3200+ CG runs 250mhz at 2.5ghz. there for your 3000+ is only running,, what ?? 2.3ghz ??

so recheck which chip you have cuz if its a 10x multiplier then you have a 3200+ not a 3000+
Click to expand...
he has a 2.0ghz newcastle with 512kb of cache, not a 1.8ghz clawhammer with 1mb


v_lestat said:
what makes me so mad about my overclock is that im stuck at 2.5ghz.

nothing i do allows for a higer clockrate.
Click to expand...
be happy, i need 1.7v for just 2.4ghz, forget 2.5 without water cooling...
 
v_lestat said:
that info is not possible.
if its a 3000+ then
#1 - the mulitplier is infact 9 not 10. or is it 8 ?
#2 - a FSB setting of 250mhz does NOT make 2.5ghz - my 3200+ CG runs 250mhz at 2.5ghz. there for your 3000+ is only running,, what ?? 2.3ghz ??

so recheck which chip you have cuz if its a 10x multiplier then you have a 3200+ not a 3000+
Click to expand...

The 3000s run at 2ghz stock, 200 FSB, ie a 10 multiplier ('least the Newcastle socket 754 3000s do, s939 run at 1.8ghz stock). If your 3200 runs at 2ghz then it's probably just a Clawhammer core which had more L2 cache, thus a lower stock clock. All 3000s are Newcastles I think though, with half the cache of a Clawhammer. Eitehr way, all current s754 3000s definitely use a stock multiplier of 10.
 
v_lestat said:
what makes me so mad about my overclock is that im stuck at 2.5ghz.

nothing i do allows for a higer clockrate.

i have NOT however dropped the multiplier down to 9 or 8 and check to see what i can crank the fsb up to.
simply because the mofo's at asus only allow for vdimm voltages of 2.7v MAX,, thats the most stupid thing ive seen a mobo maker do and most suprising is that asus did it.. those idiots should now better.

anyone know if the asus K8N-E is doing anybetter than what my sig says ?
i am thinking of getting one of the ocz ddr boosters but i fear i may waste my money if i cant get a better overclock.
Click to expand...
Hey Lestat, check this out:http://xtremesystems.org/forums/showthread.php?t=41238
 
Impulse said:
The 3000s run at 2ghz stock, 200 FSB, ie a 10 multiplier ('least the Newcastle socket 754 3000s do, s939 run at 1.8ghz stock). If your 3200 runs at 2ghz then it's probably just a Clawhammer core which had more L2 cache, thus a lower stock clock. All 3000s are Newcastles I think though, with half the cache of a Clawhammer. Eitehr way, all current s754 3000s definitely use a stock multiplier of 10.
Click to expand...



i wasnt aware of that,,, thank you and my apologies for being a dumbass ya learn something everyday


oh and noway im voltmodding this board.
ill spend the 35 bucks on an OCZ Booster.
i just dont want to cuz asus needs to pull thier head out of their ass and update the bios to allow more voltages
 
S754 CH @ 2.5GHz? Unless you have water cooling or phase change, that's pretty good for air.
 
2800+ (1.8Ghz) @ 2.43 9x270 on air (Zalman Cau) @ 1.6 Volts (S754)

never goes above 53 C full load ;)
 
the 2.5ghz was the easy part,, its rock solid, earlier today i dropped the multiplier to 9 and went to 260 and it held there.. so i raised it back to 10x. and 255mhz. for a clock of 2.55 so ill jump to 260 and see if she holds.

and i was running 250mhz on air with heat never going above 45 full load.
but now im on water with average temp running 32 deg and not higher than 38. and thats only cuz it was really warm in here last night for some wierd reason. but right now its 34.
video card thats in the sam water loop is 48 idle and like 65 something load. whcih is a nice temp for a BFG 6800GT OC at load.

ok anyways. im pretty sure i could do better if i could throw some more voltage to the ram but i cant be sure. its worth a shot. i just dont wanna spend the money for the ocz booster but i think im gunna have to. now if i could get some peeps to buy some stuff from my fs/ft thread,, i could get the booster stick (sorry for the shameless plug).
 
How old is your OCZ EL PC4000? The older stuff has Hynix D43, which will reach the point of diminishing returns at around 2.7-2.8V.
 
only about 2 months old the rev 2 stuff

yeah system wont boot past 255mhz fsb
at 258 the windows loading comes up and stays there. with the little bar going back and forth and at 260 the system dont make it to the windows loading screen and reboots.
 
I'm getting 260mhz x 10 = 2600 , and 4x HTT on my 3200+ 90nm winchester. I'm using air cooling with the Thermalright XP-120 on a MSI K8N-NEO2-Platinum/54 (yes it fits with a little bending). My memory is 2x 512mb Crucial ballistic PC4000.

Been running for 4 days now , seems entirely stable , played thru most of halflife 2 without any problems.

It boots into windows at 2700mhz but crashes on sisoft sandra, im only giving it 1.60vcore thou.
 
Havent tried lower htt , thought that maybe going lower than stock HTT would impair on performance. But beeing new to this platform i havent tried too much yet. You think lowering HTT multiplier might make 270mhz fsb stable ?
 
is 800 I believe - 1000 in the ultra and/or NF4 ??

Mine is at 270 FSB and 3X htt for 810 runs rock solid.
 
hey notoriousformula, what kinda voltage you feeding that ram? and what about system temps? I've got a very similar setup and wondering how far to push it.

So far - A64 3000 @ 2600mhz w/ 1.65v
OCZ 3200 Platin2 @ 260mhz w/3.0v @ 2.5-4-3-10

running 44deg idle and 55deg max load

Pretty sure I can go more, but the temps are scaring me off. I know temp readings are famously inaccurate, but I still need some consolation. Anybody else have a Zalman copper cooler? Maybe I should remount it. . .
 
Optikalusion said:
running 44deg idle and 55deg max load

Pretty sure I can go more, but the temps are scaring me off. I know temp readings are famously inaccurate, but I still need some consolation. Anybody else have a Zalman copper cooler? Maybe I should remount it. . .
Click to expand...
i've got that zalman.. it takes a while to heat up, but once it does.. it's not budgin any higher. the highest i've ever been able to get mine to go is 60ºc at 2.5ghz and 1.85v.. too bad i can't run high volts through my chip all the time.. *shrug*
what board do you have though? i don't really recall anything about any boards misreading low.. it's always high, so you should be safe for another 10º or so.
 
Status
Not open for further replies.
Back
Top